The long term objective of the proposed research is to understand how cochlear implant patients recognize speech when presented by a few channels of stimulation. To achieve this objective we propose converging experiments with normal-hearing listeners and cochlear-implant patients. Our experiments fall into three, broad categories. In the first category are four experiments in which the aim is to compare the performance of normally hearing subjects listening to speech processed in the manner of an implant processor with the performance of implant patients. Our goal is to assess the reality of implant performance in quiet and in noise against the potential level of performance which might be achieved with implant systems. In the second category of experiments our aim, in seven experiments, is to assess the effects of selected signal-processor and electrode parameters on speech understanding. Our goal is to uncover processor parameters which will enhance speech understanding. We will assess speech understanding under conditions in which the signal processing differs, e.g., fixed channel processors vs. n-of-m (SPEAK-like) processors, and CIS vs.SAS processors; conditions in which the spacing and apical extent of electrodes differ; and conditions in which the algorithms which map input signals on the compression function differ. In the third category of experiments is a single experiment in which we model the use of channel amplitudes by implant patients for vowel recognition. Overall, our experiments are designed, first, to assess how well implants function relative to how well they could function; second, to determine processor variables which could enhance patient performance; and, third, to provide a deeper understanding of how patients use the information delivered by a signal processor to recognize the sounds of speech.